Dual Band Thermal Videography: Separating Time-Varying Reflection and Emission Near Ambient Conditions

Abstract

Long-wave infrared radiation captured by a thermal camera includes (a) emission from an object governed by its temperature and emissivity, and (b) reflected radiation from the surrounding environment. Separating these components is a long-standing challenge in thermography. Even when using multiple bands, the problem is under-determined without priors on emissivity. This difficulty is amplified in near ambient conditions, where emitted and reflected signals are of comparable magnitude. We present a dual-band thermal videography framework that reduces this ambiguity by combining two complementary ideas at a per-pixel level: (i) spectral cues (ratio of emissivity between bands is unknown but fixed), and (ii) temporal cues (object radiation changes smoothly while background radiation changes rapidly). We derive an image formation model and an algorithm to jointly estimate the object's emissivity at each band, and the time-varying object and background temperatures. Experiments with calibrated and uncalibrated emissivities in everyday scenes (e.g., coffee pot heating up, palm print on mirrors, reflections of moving people), demonstrate robust separation and recovery of temperature fields.

Figures (10)

• Figure 1: Image formation in a thermal camera comprises of radiation from the object $\Phi_s$, reflection $\Phi_b$, optics $\Phi_o$, transmission path $\Phi_t$ and its internal components $\Phi_i$.
• Figure 2: [Left] Comparison of our method with a recent BCP Grimming2023 technique, naive multi-wavelength, and dual-wavelength pyrometry araujo2017multidualWavelengthPryo on simulated thermal videos of different materials sourced from the spectral library ECOSTRESSv1ECOSTRESSv2. The naive least squares result uses the best of five initializations selected based on the one that achieved the least objective. All methods degrade at high noise; ours improves significantly at moderate noise (log scale). See our plot in the supplementary for per-material comparisons. [Right] Shows the error in estimated emissivities from our method for a range of emissivities under varying noise levels in simulation.
• Figure 3: Spatial temperature errors (in $^\circ C$) averaged over video frames for our method and naive least squares (LS). The thermal video (left) was obtained through a combination of heat simulation and a moving toy background based on Eq. \ref{['eq:multi_band_LT']} with added noise.
• Figure 4: [Top] Images captured with different spectral filters used for calibration. The red bounding box highlights the reflected blackbody on the object. A thermocouple is placed next to this reflection to measure the object's temperature. [Bottom] Calibrated emissivity values for various materials using our technique align the closely with reported values design1st_emissivity_valuesECOSTRESSv2 as shown in Tab. \ref{['tab:emiss_comp']}.
• Figure 5: Finger print versus Finger reflection: [Top] Separating finger prints on a glass plate that emit light (heat transport) from the reflection of the uniform background (light transport). [Bottom] Separating the reflection of the fingers (light transport) by the glass plate at constant room temperature (heat transport).